[lustre-devel] [PATCH] staging: lustre: ldlm: use designated initializers

[Patrick Farrell <paf@cray.com>]

James,


This should be a purely syntactical change, to help out tools - for GCC, I'm pretty sure the meaning of { } and { NULL } are the same.  Also, I don't think struct randomization does what you're thinking.


Kees,


Is there anything written up on kernel struct randomization?  I was trying to find a talk/post from you or something from LWN, but I couldn't find something about this specifically.  (Probably because I can't find it among the other stuff that's been written up)


- Patrick

________________________________
From: lustre-devel <lustre-devel-bounces@lists.lustre.org> on behalf of James Simmons <jsimmons@infradead.org>
Sent: Monday, December 19, 2016 10:22:58 AM
To: Kees Cook
Cc: devel at driverdev.osuosl.org; Greg Kroah-Hartman; linux-kernel at vger.kernel.org; Oleg Drokin; Vitaly Fertman; Bruce Korb; Emoly Liu; lustre-devel at lists.lustre.org
Subject: Re: [lustre-devel] [PATCH] staging: lustre: ldlm: use designated initializers


> Prepare to mark sensitive kernel structures for randomization by making
> sure they're using designated initializers. These were identified during
> allyesconfig builds of x86, arm, and arm64, with most initializer fixes
> extracted from grsecurity.
>
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
>  drivers/staging/lustre/lustre/ldlm/ldlm_flock.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
> index 722160784f83..f815827532dc 100644
> --- a/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
> +++ b/drivers/staging/lustre/lustre/ldlm/ldlm_flock.c
> @@ -143,7 +143,7 @@ static int ldlm_process_flock_lock(struct ldlm_lock *req, __u64 *flags,
>        int added = (mode == LCK_NL);
>        int overlaps = 0;
>        int splitted = 0;
> -     const struct ldlm_callback_suite null_cbs = { NULL };
> +     const struct ldlm_callback_suite null_cbs = { };
>
>        CDEBUG(D_DLMTRACE,
>               "flags %#llx owner %llu pid %u mode %u start %llu end %llu\n",

Nak. Filling null_cbs with random data is a bad idea. If you look at
ldlm_lock_create() where this is used you have

if (cbs) {
        lock->l_blocking_ast = cbs->lcs_blocking;
        lock->l_completion_ast = cbs->lcs_completion;
        lock->l_glimpse_ast = cbs->lcs_glimpse;
}

Having lock->l_* point to random addresses is a bad idea.
What really needs to be done is proper initialization of that
structure. A bunch of patches will be coming to address this.
_______________________________________________
lustre-devel mailing list
lustre-devel at lists.lustre.org
http://lists.lustre.org/listinfo.cgi/lustre-devel-lustre.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-devel-lustre.org/attachments/20161219/bd5785da/attachment.htm>